An instance of this class will define metadata and data storage provider for gathering telemetry analytics of a Shiny dashboard.
The name and version parameters will describe the dashboard name and
version to track using analytics, allowing to store the analytics data from
multiple dashboards in the same data storage provider. As well as
discriminate different versions of the dashboard.
The default data storage provider uses a local SQLite database, but this
can be customizable when instantiating the class, by using another one of
the supported providers (see DataStorage).
Debugging
Events are logged at the DEBUG level using the logger package.
To see the logs, you can set:
logger::log_threshold("DEBUG", namespace = "shiny.telemetry")See also
DataStorage which this function wraps.
Active bindings
data_storageinstance of a class that inherits from
DataStorage. See the documentation on that class for more information.app_namestring with name of dashboard
Methods
Method new()
Constructor that initializes Telemetry instance with parameters.
Usage
Telemetry$new(
app_name = "(dashboard)",
data_storage = DataStorageSQLite$new(db_path = file.path("telemetry.sqlite"))
)Arguments
app_name(optional) string that identifies the name of the dashboard. By default it will store data with
(dashboard).data_storage(optional)
DataStorageinstance where telemetry data is being stored. It can take any of data storage providers by this package, By default it will store in a SQLite local database in the current working directory with filenametelemetry.sqliteversion(optional) string that identifies the version of the dashboard. By default it will use
v0.0.0.
Method start_session()
Setup basic telemetry
Usage
Telemetry$start_session(
track_inputs = TRUE,
track_values = FALSE,
login = TRUE,
logout = TRUE,
browser_version = TRUE,
navigation_input_id = NULL,
session = shiny::getDefaultReactiveDomain(),
username = NULL,
track_anonymous_user = TRUE,
track_errors = TRUE
)Arguments
track_inputsflag that indicates if the basic telemetry should track the inputs that change value.
TRUEby defaulttrack_valuesflag that indicates if the basic telemetry should track the values of the inputs that are changing.
FALSEby default. This parameter is ignored iftrack_inputsisFALSEloginflag that indicates if the basic telemetry should track when a session starts.
TRUEby default.logoutflag that indicates if the basic telemetry should track when the session ends.
TRUEby default.browser_versionflag that indicates that the browser version should be tracked.
TRUEby default.navigation_input_idstring or vector of strings that represent input ids and which value should be tracked as navigation events. i.e. a change in the value represent a navigation to a page or tab. By default, no navigation is tracked.
sessionShinySessionobject or NULL to identify the current Shiny session.usernameCharacter with username. If set, it will overwrite username from session object.
track_anonymous_userflag that indicates to track anonymous user. A cookie is used to track same user without login over multiple sessions, This is only activated if none of the automatic methods produce a username and when a username is not explicitly defined.
TRUEby default.track_errorsflag that indicates if the basic telemetry should track the errors.
TRUEby default. if using shiny version <1.8.1, it can auto log errors only in UI output functions. By using latest versions of shiny, it can auto log all types of errors.
Method log_navigation()
Log an input change as a navigation event
Usage
Telemetry$log_navigation(input_id, session = shiny::getDefaultReactiveDomain())Method log_navigation_manual()
Log a navigation event manually by indicating the id (as input id)
Usage
Telemetry$log_navigation_manual(
navigation_id,
value,
session = shiny::getDefaultReactiveDomain()
)Method log_login()
Log when session starts
Usage
Telemetry$log_login(
username = NULL,
session = shiny::getDefaultReactiveDomain()
)Method log_logout()
Log when session ends
Usage
Telemetry$log_logout(
username = NULL,
session = shiny::getDefaultReactiveDomain()
)Method log_click()
Log an action click
Usage
Telemetry$log_click(id, session = shiny::getDefaultReactiveDomain())Method log_browser_version()
Log the browser version
Usage
Telemetry$log_browser_version(session = shiny::getDefaultReactiveDomain())Method log_button()
Track a button and track changes to this input (without storing the values)
Usage
Telemetry$log_button(
input_id,
track_value = FALSE,
session = shiny::getDefaultReactiveDomain()
)Arguments
input_idstring that identifies the button in the Shiny application so that the function can track and log changes to it.
track_valueflag that indicates if the basic telemetry should track the value of the input that are changing.
FALSEby default.sessionShinySessionobject or NULL to identify the current Shiny session.
Method log_all_inputs()
Automatic tracking of all input changes in the App. Depending on the
parameters, it may only track a subset of inputs by excluding patterns
or by including specific vector of input_ids.
Usage
Telemetry$log_all_inputs(
track_values = FALSE,
excluded_inputs = c("browser_version"),
excluded_inputs_regex = NULL,
include_input_ids = NULL,
session = shiny::getDefaultReactiveDomain()
)Arguments
track_valuesflag that indicates if the basic telemetry should track the values of the inputs that are changing.
FALSEby default. This parameter is ignored iftrack_inputsisFALSE.excluded_inputsvector of input_ids that should not be tracked. By default it doesn't track browser version, which is added by this package.
excluded_inputs_regexvector of input_ids that should not be tracked. All Special characters will be escaped.
include_input_idsvector of input_ids that will be tracked. This input_ids should be an exact match and will be given priority over exclude list.
sessionShinySessionobject or NULL to identify the current Shiny session.
Method log_input()
Track changes of a specific input id.
Usage
Telemetry$log_input(
input_id,
track_value = FALSE,
matching_values = NULL,
input_type = "text",
session = shiny::getDefaultReactiveDomain()
)Arguments
input_idstring (or vector of strings) that identifies the generic input in the Shiny application so that the function can track and log changes to it.
When the
input_idis a vector of strings, the function will behave just as callinglog_inputone by one with the same arguments.track_valueflag that indicates if the basic telemetry should track the value of the input that are changing.
FALSEby default.matching_valuesAn object specified possible values to register.
input_type"text"to registered bare input value,"json"to parse value fromJSONformat.sessionShinySessionobject or NULL to identify the current Shiny session.
Method log_input_manual()
Log a manual input value.
This can be called in telemetry and is also used as a layer between log_input family of functions and actual log event. It creates the correct payload to log the event internally.
Usage
Telemetry$log_input_manual(
input_id,
value = NULL,
session = shiny::getDefaultReactiveDomain()
)Method log_custom_event()
Log a manual event
Usage
Telemetry$log_custom_event(
event_type,
details = NULL,
session = shiny::getDefaultReactiveDomain()
)Method log_error()
Log a manual error event
Usage
Telemetry$log_error(
output_id,
message,
session = shiny::getDefaultReactiveDomain()
)Method log_errors()
Track errors as they occur in observe and reactive
Usage
Telemetry$log_errors(session = shiny::getDefaultReactiveDomain())Examples
log_file_path <- tempfile(fileext = ".txt")
telemetry <- Telemetry$new(
data_storage = DataStorageLogFile$new(log_file_path = log_file_path)
) # 1. Initialize telemetry with default options
#
# Use in a shiny application
if (interactive()) {
library(shiny)
shinyApp(
ui = fluidPage(
use_telemetry(), # 2. Add necessary javascript to Shiny
numericInput("n", "n", 1),
plotOutput('plot')
),
server = function(input, output) {
telemetry$start_session() # 3. Minimal setup to track events
output$plot <- renderPlot({ hist(runif(input$n)) })
}
)
}
#
# Manual logging of Telemetry that can be used inside Shiny Application
# to further customize the events to be tracked.
session <- shiny::MockShinySession$new() # Create dummy session (only for example purposes)
class(session) <- c(class(session), "ShinySession")
telemetry$log_click("a_button", session = session)
telemetry$log_error("global", message = "An error has occured")
telemetry$log_custom_event("a_button", list(value = 2023), session = session)
telemetry$log_custom_event("a_button", list(custom_field = 23), session = session)
# Manual call login with custom username
telemetry$log_login("ben", session = session)
# Read all data
telemetry$data_storage$read_event_data()
#> # A tibble: 5 × 11
#> app_name type session id output_id message value custom_field username
#> <chr> <chr> <chr> <chr> <chr> <chr> <chr> <chr> <chr>
#> 1 (dashboard) click 6df7df… a_bu… NA NA NA NA NA
#> 2 (dashboard) error NA NA global An err… NA NA NA
#> 3 (dashboard) a_but… 6df7df… NA NA NA 2023 NA NA
#> 4 (dashboard) a_but… 6df7df… NA NA NA NA 23 NA
#> 5 (dashboard) login 6df7df… NA NA NA NA NA ben
#> # ℹ 2 more variables: time <dttm>, date <date>
file.remove(log_file_path)
#> [1] TRUE
#
# Using SQLite
db_path <- tempfile(fileext = ".sqlite")
telemetry_sqlite <- Telemetry$new(
data_storage = DataStorageSQLite$new(db_path = db_path)
)
telemetry_sqlite$log_custom_event("a_button", list(value = 2023), session = session)
telemetry_sqlite$log_custom_event("a_button", list(custom_field = 23), session = session)
# Read all data from time range
telemetry_sqlite$data_storage$read_event_data("2020-01-01", "2055-01-01")
#> # A tibble: 2 × 9
#> time app_name session type value custom_field date id
#> <dttm> <chr> <chr> <chr> <chr> <chr> <date> <chr>
#> 1 2024-10-17 10:47:49 (dashbo… 6df7df… a_bu… 2023 NA 2024-10-17 NA
#> 2 2024-10-17 10:47:49 (dashbo… 6df7df… a_bu… NA 23 2024-10-17 NA
#> # ℹ 1 more variable: username <chr>
file.remove(db_path)
#> [1] TRUE